Correctly process compile.disable in config

[MGAMZ]

This is a sub-PR of open-mmlab#1665

Motivation

The torch.compile receives a lot args. And in mmengine, all contents in compile from any config file will be transferred to torch.compile.
The mmengine checks several dependencies when user specifies compile in their config file, and making hasattr(config, compile) as the flag of enabling torch.compile.
However this is not exact. The compile config can include an arg disable to state if the compiler is actually working. So the mmengine needs to carefully determine if the user is actually enabling the torch.compile. This caused a small modification in mmengine/_strategy/base.py

Modification

When user specified compile with a dict type, the compile_model func will check if disable is set to True, and can directly return the model just like compile is set to None

[read-the-docs-community]

Documentation build overview

📚 onedl-mmengine | 🛠️ Build #30160452 | 📁 Comparing d93517e against latest (a7f271e)

🔍 Preview build

Show files changed (5 files in total): 📝 5 modified | ➕ 0 added | ➖ 0 deleted

File	Status
advanced_tutorials/data_transform.html	📝 modified
advanced_tutorials/test_time_augmentation.html	📝 modified
migration/transform.html	📝 modified
api/generated/mmengine.model.BaseTTAModel.html	📝 modified
_modules/mmengine/_strategy/base.html	📝 modified

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.